(179g) Predicting Activation Energies of Transition Metal Nanoparticle Reconstruction Via an Interaction-Counting Approach

Reducing the size of catalyst materials is a key to enhancing their atomic utilization efficiency, particularly important for expensive transition metal catalysts. As the size of the catalyst reduces, the high surface energy associated with nanofeatured structures imparts a thermodynamic driving force for atomic-scale restructuring. Understanding the dynamic nature of nanoparticles is crucial when designing stable catalysts with atomic-scale precision; it is also important for identifying potential catalytically active transient states. However, the myriad geometric structures available during nanoparticle reconstruction, along with the high computational cost for evaluating related energetics (thermochemistry and kinetics), makes it critical to develop an efficient and accurate method for representing the energetics associated with nanoparticle reconstruction if the potential of computational predictive power is to be fully realized.

In this presentation, we will discuss a DFT-based scheme for predicting activation energies of atomic processes involved in transition metal reconstruction using an interaction-counting approach inspired by recent work describing the thermochemistry of varying metal arrangements.¹ Based on a training set of activation energies explicitly calculated on simple surface slab models, we predict activation energies for more complex diffusion events including (i) hopping events along edges and over steps, and (ii) substitutional events into edges, steps, and terraces of seven transition metals (Cu, Rh, Pd, Ag, Ir, Pt, and Au). We apply this method to predict energetics of analogous processes on nanoparticle models, finding that our approach is sufficiently accurate (mean error ~ 0.1 eV) to enable future application in stochastic modeling techniques (e.g., kinetic Monte Carlo²).

References

1. Roling, L. T. et al. J. Phys. Chem. C., 121, 41, 23002-23010, 2017.

2. Liu, D. J. et al. Chem. Rev., 115, 12, 5979-6050, 2015.